Content distributing algorithm for replicated Web servers based on market models

A content distributing algorithm based on market models for a replication distribution scheme of Web contents is proposed. In general, replication distribution is considered to effectively utilize the time and space local properties of access from a client and has been the topic of numerous studies....

Full description

Saved in:
Bibliographic Details
Published inElectronics & communications in Japan. Part 2, Electronics Vol. 84; no. 11; pp. 9 - 19
Main Authors Nishimura, Kenji, Yamamoto, Miki, Ikeda, Hiromasa
Format Journal Article
LanguageEnglish
Published New York John Wiley & Sons, Inc 01.11.2001
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:A content distributing algorithm based on market models for a replication distribution scheme of Web contents is proposed. In general, replication distribution is considered to effectively utilize the time and space local properties of access from a client and has been the topic of numerous studies. However, maintaining QoS (such as shortening the content obtaining time) in replicated servers when servers are in a congested state has not been considered in these studies. Since replicated servers constitute a resource shared by multiple ISPs (Internet Service Providers), the accesses need to be limited to replicated servers in case of a congested state to maintain the QoS enjoyed by each client. Thus, in this paper, a scheme for determining allocation of access to replicated servers for each content during congestion based on market models is proposed. The analytical and simulation results have confirmed that the proposed method controls the QoS in accordance with the access arrival rate to each content. © 2001 Scripta Technica, Electron Comm Jpn Pt 2, 84(11): 9–19, 2001
Bibliography:ark:/67375/WNG-1T8PZLC9-R
ArticleID:ECJB1068
istex:0803D1BBA1F10AF7572CD1428D1FEBBFDD461CE4
Scientific Research Fund, Ministry of Education
ObjectType-Article-2
SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 23
ISSN:8756-663X
1520-6432
DOI:10.1002/ecjb.1068